Neural network modelling of constrained spatial information flows: Design, estimation and performance issues

Neural network modelling of constrained spatial information flows: Design, estimation and performance issues

Abstract

In this chapter a novel modular product unit neural network architecture is presented to model singly constrained spatial interaction flows. The efficacy of the model approach is demonstrated for the origin-constrained case of spatial interaction using Austrian interregional telecommunication traffic data. The model requires a global search procedure for parameter estimation, such as the Alopex procedure. A benchmark comparison against the standard origin-constrained gravity model and the two-stage neural network approach, suggested by Openshaw (1998), illustrates the superiority of the proposed model in terms of the generalisation performance measured by ARV and SRMSE.

Grafik Top
Authors
  • Fischer, Manfred M.
  • Reismann, Martin
  • Hlavackova-Schindler, Katerina
Grafik Top
Shortfacts
Category
Book Section/Chapter
Divisions
Data Mining and Machine Learning
Subjects
Kuenstliche Intelligenz
Title of Book
Spatial Analysis and GeoComputation: Selected Essays
ISSN/ISBN
978-3-540-35730-8
Page Range
pp. 241-268
Date
January 2006
Official URL
http://link.springer.com/chapter/10.1007/3-540-357...
Export
Grafik Top